On Surgical Treatment of Aortic Pathology

Full text

(1)Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1613. On Surgical Treatment of Aortic Pathology JACOB BUDTZ-LILLY. ACTA UNIVERSITATIS UPSALIENSIS UPPSALA 2019. ISSN 1651-6206 ISBN 978-91-513-0806-7 urn:nbn:se:uu:diva-395964.

(2) Dissertation presented at Uppsala University to be publicly examined in Rosénsalen, entrance 95/96, NBV, Akademiska sjukhuset, Uppsala, Wednesday, 18 December 2019 at 13:00 for the degree of Doctor of Philosophy (Faculty of Medicine). The examination will be conducted in English. Faculty examiner: Associate Professor Philip Goodney (Dartmouth Institute). Abstract Budtz-Lilly, J. 2019. On Surgical Treatment of Aortic Pathology. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1613. 61 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0806-7. The use of endovascular aneurysm repair (EVAR) in the treatment of abdominal aortic aneurysms has advanced from a premature characterization as a “failed experiment” in early 2000 to the predominant modern method of treatment. Technology has accommodated initial shortcomings, but it has also led to expansions in the treatment of ruptured aneurysms and complex aortic pathologies. The overall aim of this thesis is to characterize the contemporary utilization of endovascular repair in the international setting and to evaluate its expanding use in complex aortic disease treatment. Paper I is an analysis of outcomes after intact aneurysm treatment from registries of 12 countries. From 2005 to 2013, and with 83,253 patients included, it was shown that the use of EVAR has increased while, the perioperative mortality has decreased. This was counterbalanced by a worsening mortality for those patients treated with open aortic repair. Paper II is an analysis of ruptured aneurysms from the above-mentioned international registries. EVAR is also increasing for these patients, although open repair is still the predominant treatment strategy in most centres. Perioperative mortality was superior for EVAR patients, despite increased age and comorbidities. An association between patient-volume and perioperative mortality could be shown for open repair, but the same could not be demonstrated for EVAR. Paper III is an evaluation of the adaptation of a total endovascular approach for the treatment of complex abdominal aortic aneurysms from a single centre. The technical success and midterm mortality, as well as post-operative complications, including spinal ischemia, were similar to those reported from large and multi-centre analyses. Previous studies reveal disparate results for centres performing open complex aortic repair. The results here suggest that a total endovascular approach is feasible for dedicated centres contemplating this strategy. Paper IV is an analysis of multiple pre-, peri-, and post-operative variables documented from complex aneurysm procedures. A relationship between increased complexity and variables such as anaesthesia duration, bleeding, hospital stay, and radiation exposure was found. As patients and their comorbidities increase, a decision to embark on a complex procedure should be made with due diligence to these relationships. Paper V is a technical analysis of patients following acute treatment for Type A aortic dissections. Many patients are unfit for open aortic arch repair. Based on current availability of endovascular aortic stentgrafts, it was shown that the majority of patients can be treated endovascularly, while anticipated device improvements should further increase the proportion of eligibility. Keywords: Abdominal aortic aneurysm, EVAR, F/BEVAR, aorta dissection. Jacob Budtz-Lilly, Department of Surgical Sciences, Vascular Surgery, Akademiska sjukhuset ing 70 1 tr, Uppsala University, SE-751 85 Uppsala, Sweden. © Jacob Budtz-Lilly 2019 ISSN 1651-6206 ISBN 978-91-513-0806-7 urn:nbn:se:uu:diva-395964 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-395964).

(3) This work is more than a dedication to my family. It could not exist without your support, and nor could I, without the love and radiance of life you teach and give me throughout. Thank you Anna, Ellen, Thorleif, and Frej Lucas..

(4) Those among us who are unwilling to expose their ideas to the hazard of refutation do not take part in the scientific game. -Karl Raimund Popper.

(5) List of Papers. This thesis is based on the following papers, which are referred to in the text by their Roman numerals. I. Budtz-Lilly J, Venermo M, Debus S, Behrendt CA, Altreuther M, Beiles B, et al. (2017) Editors’s Choice-Assessment of International Outcomes of Intact Abdominal Aortic Aneurysm Repair over 9 years. Eur J Vasc Endovasc Surg, 54(1):13-20.. II. Budtz-Lilly J, Björck M, Venermo M, Debus S, Behrendt CA, Altreuther M, et al. (2018) Editor’s Choice-The Impact of Centralisation and Endovascular Aneurysm Repair on Treatment of Ruptured Abdominal Aortic Aneurysms Based on International Registries. Eur J Vasc Endovasc Surg, 56(2):181-88.. III. Budtz-Lilly J, Wanhainen A, Erikson J, Mani K. (2017) Adapting to a total endovascular approach for complex aortic aneurysm repair: Outcomes after fenestrated and branched endovascular aortic repair. J Vasc Surg, 66(5):1349-56.. IV. Budtz-Lilly J, Liungman K, Wanhainen A, Mani K. (2019) Correlations Between Branch Vessel Catheterization and Procedural Complexity in Fenestrated and Branched Endovascular Aneurysm Repair. Vasc Endovascular Surg, 53(4):277-83.. V. Budtz-Lilly J, Vikholm P, Wanhainen A, Astudillo R, Thelin S, Mani K. (2019) Technical eligibility for endovascular treatment of the aortic arch after open type A aortic dissection repair.. Reprints were made with permission from the respective publishers..

(6)

(7) Contents. Introduction Historical Aspects and Connotations Epidemiology and Definitions Intact AAA Repair Ruptured AAA Repair The Complex Aortic Repair Outcomes and Registry Analysis. 11 11 12 13 13 14 17. Aims. 19. Patient and Methods Papers I and II Papers III and IV Paper V Statistics. 20 20 21 22 25. Ethical Considerations. 27. Results Paper I Paper II Paper III Paper IV Paper V. 28 28 30 35 39 41. General Discussion. 45. Future Perspectives. 51. Acknowledgements. 52. References. 54.

(8)

(9) Abbreviations. AAA AAD AIBS ANOVA BEVAR CABG CI CT DM EVAR EVAR(p) EVAS FEVAR Gy IA IBD ICU J/PAAA LCC LSA NICE OAR OAR(p) OR PTA RAAA RCC TAAA TEVAR. Abdominal aortic aneurysm Stanford Type A Aortic Dissection Arch Inner-Branched Stentgraft Analysis of variance Branched endovascular aortic aneurysm repair Coronary artery bypass graft Confidence interval Computed tomography Diabetes mellitus Endovascular aortic repair Predominantly EVAR Endovascular aneurysm sealing Fenestrated endovascular aortic aneurysm repair Gray Innominate artery Iliac branched device Intensive Care Unit Juxta/Pararenal abdominal aortic aneurysm repair Left common carotid artery Left subclavian artery National Institute for Health and Care Excellence Open aortic repair Predominantly OAR Odds ratio Percutaneous transluminal angioplasty Ruptured abdominal aortic aneurysm Right common carotid artery Thoracoabdominal aortic aneurysm Thoracic endovascular aortic repair.

(10)

(11) Introduction. Historical Aspects and Connotations “An aneurysm is the dilatation of an artery full of spirituous blood” wrote Fernilius.1 A both pithy and scientific observation, such as this, resonates well with the still present fear that the ruptured aneurysm connotes to this day. A dilatation or “widening”, as the Greek term “aneurysm” defines, has been recognized as a blood vessel pathology spanning two millennia. Galen, a surgeon of ancient Rome, warned against the tampering or wounding of aneurysms, for “the blood is spouted out with so much violence that it can scarcely be arrested.”2 Even centuries later, little waning of this distress is noted; caution and restraint prevailed. As one patient of the anatomist William Hunter consoled himself, “You know, Mr Hunter’s advice to me was, not to do anything to it; which I have satisfied myself with…”3 Trepidation notwithstanding, efforts in treating patients with aneurysms were not forfeited. The more sanguine Hunter brother, John, experimented and successfully treated some patients with popliteal aneurysms by proximal vessel ligation.4 Albeit a milestone in the surgical historical records, it should be recalled that many of Hunter’s colleagues, including Percival Pott, considered ligation as inadequate. Astley Cooper later echoed this discountenance in 1817, as he attempted to ligate the aorta for a ruptured iliac aneurysm, for which the patient died 40 hours later.5 It was not until the twentieth century that substantial ground was attained in treating aneurysms of the abdominal aorta. The concept of endoaneurysmorrhaphy, or preservation of the aortic lumen with obliteration of the aneurysmal sac and collateral vessel ligation, was advanced successfully by Rudolf Matas.6 One of his contemporaries, Arthur Voorhees, advocated for and developed the use of synthetic aortic prostheses in order to reestablish lower body blood flow, once the aneurysm itself was removed.7 One of his prostheses was used in 1952 for a patient with a ruptured AAA. Open surgical repair of abdominal aortic aneurysms, both intact and ruptured, thus became more widespread. Its grave connotations with acute surgery did not abate, however, as well-known persons such as Albert Einstein, underwent AAA treatment. Interestingly, Einstein received an alternative repair, one involving cellophane wrapping of the aneurysm to prevent rupture, which proved to be inadequate.8 When his aneurysm indeed ruptured several years later, having been offered an acute reoperation, he provided a sobering 11.

(12) reminder to the vascular surgeons, “I want to go when I want. It is tasteless to prolong life artificially.” The advent of catheter-based blood vessel repair and interventional radiology were milestones in the history of medicine.9 Sven Ivar Seldinger was instrumental in refining vascular access, using a guidewire through a needle and thereby minimizing trauma, an eponymous technique still used today.10 Coupled with radiographic imaging, this type of minimally invasive therapy was rapidly developed in treating atherosclerotic diseased vessels. The Ukraninan surgeon Nikolai Volodos, in 1984, was the first to successfully use an endoprosthesis within an AAA.11,12 Juan Parodi refined this technique by 1990, whereby he placed a custom-designed Dacron tube within an aneurysm, using a transfemoral approach.13 Today’s hindsight can deceive one into believing that the subsequent rapid increase in endovascular aortic repair (EVAR) was inevitable and welcome. It is hazardous to generalize its reception by the vascular surgery community, but the British Journal of Surgery editorial on EVAR in 2001, subtitled, “A Failed Experiment” does provide some parity to the presupposed enthusiasm.14 Since then, there has been no lack of scientific studies to address the role that EVAR should play, while the fervor that accompanies these discussions has also not been lacking.. Epidemiology and Definitions The European Society of Vascular Surgery defines the AAA as an abdominal aorta equal to or greater than 3.0 cm in either the anterior-posterior or transverse plane.15 Screening programs have revealed an AAA prevalence of 5.1% in Denmark, 1.3 % in the United Kingdom, and 1.7% in Sweden.16–18 Several studies have indicated that this prevalence is decreasing, perhaps due to decreasing rates of smoking.18,19 Both the European Society of Vascular Surgery and the Society for Vascular Surgery in the United States recommend a threshold diameter of 5.5 cm for men.15,20 Aneurysms less than this size are considered small aneurysms. Four randomized trials have addressed the proposal of earlier intervention, and none found a benefit for either endovascular or open repair of an aneurysm < 5.5 cm.21–25 Because the risk of rupture of small aneurysms among women is roughly four times higher than it is for men, repair at a diameter of 5.0 cm for women has been warranted by several analyses.26,27. 12.

(13) Intact AAA Repair Prophylactic repair of the AAA entails a balancing of the risk of repair against the risk of rupture. Added to this is the consideration of the patient’s overall life expectancy, given various comorbidities. Contemporary treatment involves either conventional open repair with aneurysm resection and vascular reconstruction, or internal sealing with an endovascular stentgraft. Both methods have been compared against each other in four randomized control trials, Table 1.28–31 In broad terms, early outcomes favored EVAR, whereas midterm results show no significant differences. With follow-up reaching a maximum of 15.8 years, the results from the United Kingdom EndoVascular Aneurysm Repair (EVAR-1) trial revealed an inferior survival benefit for EVAR, thus raising a concern of long-term durability.32 This has recently been counterbalanced, however, by the long-term results from the Open versus Endovascular Repair of Abdominal Aortic Aneurysm (OVER) trial, in which no differences in survival were observed.33 It may be that durability of EVAR devices is now less of a concern, but surveillance is still mandatory. Table 1. Summary of 30-day mortality for four randomized controlled trials comparing EVAR and OAR for intact abdominal aortic aneurysms, based on a recent metaanalysis.28–31,34 Trial DREAM28 EVAR-129 OVER30 ACE31. Year of Publication 2004 2005 2009 2011. EVAR 30-day mortality 1.2% 1.8% 0.2% 1.3%. OAR 30-day mortality 2.9% 4.3% 1.9% 0.6%. EVAR: Endovascular aneurysm repair OAR: Open aortic repair. Ruptured AAA Repair The ruptured AAA requires acute attention and, provided the patient wishes treatment, the primary goal of intervention is to prevent death: Hence, the conventional primary outcome measure of perioperative mortality. The exacerbating hypovolemic shock, often accompanied by formidable comorbidities, place the acute RAAA operation in a dire scenario with poor prognoses. For many years, the 30-day mortality was reported as high as 50%, while contemporary analyses have tempered this risk to 30-35%.35–37 The minimal invasiveness of EVAR is incitement enough to consider EVAR in the treatment of a ruptured AAA. Moreover, the use of local anesthesia, thus avoiding the hazardous repercussions of general anesthesia in these frail patients, is considered advantageous in supporting the patient 13.

(14) through the acute phase.38 The multiple logistical preparations required for an acute EVAR, on the other hand, are somewhat limiting for some institutions, a barrier admitted to by the European Society of Vascular Surgery in their guidelines.15 In any case, initial studies revealed improved mortality rates, including the multi-centre pooled analysis by Veith et al., which showed a 30day mortality of 21.2%.39 It would seem wanted, if not feasible, to perform a randomized controlled trial, comparing EVAR and open repair, but the data have been both few and limited until recently, Table 2. A systematic review from the Cochrane Library has summarized that long-term data were lacking and more high-quality trials were needed, rendering their conclusion of no differences between the methods on relevant outcomes.40 The largest and most recent trial, the Immediate Management of Patients with Ruptured Aneurysm: Open Versus Endovascular Repair (IMPROVE) has remedied some of these shortcomings. At three years, a reduced mortality for EVAR patients was accompanied by a better quality of life and at a lower cost.41. Table 2. Mortality at 30 days from 4 randomized controlled trials comparing EVAR and OAR for ruptured AAA treatment. Trial ECAR42 IMPROVE43 AJAX44 Nottingham45. Study Period. Number of Patients. 2008-2013 2009-2013 2004-2011 2002-2004. 107 613 116 32. 30-day Mortality EVAR. OAR. 18% 35% 28% 53%. 24% 37% 29% 53%. EVAR: Endovascular aortic repair OAR: Open aortic repair. The Complex Aortic Repair The treatment of aneurysms contiguous with, or extending above, the renal arteries is a challenge. The anatomic complexity increases the technical challenge, just as many of these patients have associated increasing comorbidities. Open surgical repair has been the conventional treatment for many years, and this often entails considerable significant physiological duress-large incisions, proximal aortic clamping, in addition to greater blood loss and anesthesia duration. Perioperative mortality is not surprisingly higher than it is for standard infrarenal AAA surgery, as is the risk for the dreaded risk of spinal ischemia, often cited between 4-11%.46,47 The advancement of endovascular technology has led innovators to extend stentgrafts proximally in the aorta with steps taken to accommodate the critical 14.

(15) visceral arteries. Branched endovascular aortic repair (BEVAR) or fenestrated endovascular aortic repair (FEVAR) utilizes either branches or fenestrations, or sometimes both, to preserve blood flow to what are deemed “target vessels.” Percutaneous access at the groins or upper extremities minimizes surgical trauma, as well as blood loss and aortic cross-clamping. The risk of spinal ischemia, however, is still significant, particularly as stentgraft length and aortic coverage increase. Patency of target vessels is also a long-term issue of durability, often requiring vigilant clinical follow-up. Comparison of open and endovascular surgery is not uncomplicated. Figure 1 gives an immediate and unembellished contrast of one of many differences between open and endovascular treatment of complex aortic disease. In a systematic review and meta-analysis of elective juxtarenal AAA repairs in 2015, Rao et al found similar perioperative mortalities of 4.1% in both endovascular and open repair cohorts.48 Greenberg et al provided a retrospective assessment of their thoracic and thoracoabdominal aortic aneurysms, also finding equivalent rates of 1-year mortality of 15.6% for endovascular repair and 15.9% for open repair .49 Moreover, they found that the risk of spinal ischemia was not statistically different between endovascular and open repair. They stress, however, that selection bias was at play, as many patients deemed unfit for open surgery were relegated to an endovascular approach. It is unlikely, they opine, that a randomized control trial will be carried out in the near future, given the heterogeneity of patients and repair. Complexity has not discouraged surgeons from applying endovascular techniques even more proximally, i.e., the ascending and aortic arch, an anatomical zone conventionally reserved for open surgery. Thoracic endovascular aortic repair (TEVAR) in the descending aorta is well-established, while proximal extension requires a careful and technical accounting of the supraaortic vessels, as well as potentially limited landing zones in the ascending aorta.50 Haulon and colleagues have pioneered this development in recent years, and Verscheure et al have recently reported good results from multiple international centres of expertise.51,52 An echoing and not unfamiliar finding that has emerged is that patient and anatomical selection are essential for good results.53. 15.

(16) Figure 1a. Image from an open thoracoabdominal repair with a Dacron aortobifemoral prosthesis with four anastomosed bypasses to visceral vessels.. 16.

(17) Figure 1b. Fenestrated endovascular repair of a juxtarenal abdominal aortic aneurysm with 3 fenestrations and stentgrafts to target vessels, including the right internal iliac artery. This was performed percutaneously via bilateral groin access.. Outcomes and Registry Analysis A prospectively recorded registry of procedures and outcomes is a valuable tool for clinicians and researchers. The data reflect real-world practice, and appropriate statistical analysis can be powerful in identifying strengths and limitations of patient selection and treatment modalities. Schanzer et al, for example, demonstrated this in their analysis of AAA sac growth and the lack of adherence to instructions for use of EVAR stentgrafts.54 Although previous trials had shown the benefit of EVAR, their analysis of large registry data warned clinicians against the liberal use of EVAR in AAA treatment. Biases cannot be eliminated in registry analyses, as they are in randomized controlled trials, but they are not hampered by strict inclusion and exclusion. 17.

(18) criteria. Furthermore, they often avail themselves to either larger collations or comparisons with other registries. The value of registries is predicated on several factors. An external validation against other registries measures the inclusion of all patients and procedures and can assuage some of the selection bias. A process of internal validation evaluates the completeness of individual case records in order to insure appropriate covariable inclusion. This process of validation should then be performed regularly in order to perform robust interpretation of data.55 The data used in Papers I and II are based on the Vascunet collaboration. This international group works tirelessly to harmonize the definitions of variables and outcomes. In addition to providing benchmarks in modern vascular surgery treatment, Vascunet regularly updates documentation regarding validity of its data.56–60 The data in Papers III-IV are based on the Swedvasc registry, which is also regularly validated, as well as local surgical registries at Uppsala University Hospital.59 Unique personal identity numbers for Swedish patients furthermore allows for almost 100% follow-up data for these patients.. 18.

(19) Aims The overall aim of this thesis was to capture the contemporary outcomes and potential of endovascular aortic repair in light of the rapid expansion of treatment of intact, ruptured, and complex aortic pathologies. The specific aims were: I. To update and assess the international outcomes of intact AAA repair with a particular focus on trends in the use of EVAR vis-à-vis open repair and the implications of patient-volume on outcome.. II. To report the international trends, variations, and outcomes in the treatment of ruptured AAAs with a particular focus on the impact of EVAR and patient volumes.. III. To report on the feasibility of adapting to a total endovascular approach in the treatment of complex aortic aneurysms, specifically reporting on perioperative mortality, target-vessel stability, and complications.. IV. To describe the relationships between multiple pre-, peri-, and postoperative variables for complex aneurysm repair.. V. To characterize the eligibility of post ascending aorta dissection patients for potential retreatment using the new endovascular arch innerbranched stent graft.. 19.

(20) Patient and Methods Papers I and II Patients Patients who underwent primary repair of either an intact (Paper I) or ruptured (Paper II) AAA from 2005-2013 were included from 12 countries (Australia, Denmark, Finland, Germany, Hungary, Iceland, Italy, New Zealand, Norway, Sweden, Switzerland, and the United Kingdom). Patients who underwent reoperation were excluded. Data were obtained from the Vascunet registry, a collaboration of the above-named international vascular surgery registries. A previous Vascunet publication included data from 2005-2009, and the present analysis used these data for the purposes of comparison against data from 2010-2013. Data in the latter period were no longer available for Italy, while the addition of data from New Zealand and Iceland were now available for analysis. Methods The primary outcome for both analyses was peri-operative mortality, defined as either in-hospital death, or death within 30 days. For both papers, data were available regarding patient age and sex. When relevant, aneurysms with a reported diameter less than 5.5 cm were considered small AAAs. For Paper I, comparisons of the perioperative mortality were performed for the updated registries (2010-2013) against the previous period (2005-2009). Comprehensive data regarding comorbidities, as well as AAA diameter were absent or had varying definitions for the earlier period, thus limiting this aspect of the analysis to comparisons of age, patient sex, use of EVAR, overall perioperative mortality, and mortality based on treatment (EVAR or open repair). The strict analysis of the latter period included a comparison between countries and also included data on AAA diameter (cm). Furthermore, for the latter period, comparisons of perioperative mortality were also performed for high and low volume centres. For Paper II, only data from 2010-2013 were evaluated. Similar analyses were performed for overall perioperative mortality, as well as mortality based on type of treatment (EVAR or open repair). Further data were available for analysis, including the presence or absence of the following comorbidities: ischaemic heart disease, pulmonary disease, diabetes mellitus, and cerebrovascular disease. Patients were likewise identified with an individual centre and divided into quintiles for the purposes of comparison. Centres were also characterized as either a centre predominantly using EVAR (≥50% of RAAA. 20.

(21) repairs utilizing EVAR; EVAR(p)) as a primary strategy, or predominantly using OAR (OAR(p)) as a primary strategy.. Papers III and IV Patients All patients treated for juxta/pararenal abdominal aortic aneurysms (J/PAAA) or thoracoabdominal aortic aneurysms (TAAAs), including reoperations, from 2010-2015 from a single-centre, Uppsala University Hospital, were included. Patients with ruptured aneurysms were included in Paper III but not in Paper IV. Furthermore, any patients included in clinical trials and patients who underwent concomitant endovascular repair or iliac vessels using an iliac branched device (IBD) were excluded from Paper IV, as were any patients who underwent treatment using fusion imaging, a technology introduced at the end of this study period. All patients underwent either fenestrated endovascular aortic repair (FEVAR) or branched endovascular aortic repair (BEVAR), or a combination of these two techniques. For Paper IV, an equal number of randomly selected patients, who underwent routine AAA repair using standard EVAR, were included for comparative purposes. All devices were manufactured by Cook Medical (Bloomington, Indiana, USA). Methods The 30- and 90-day mortalities were calculated, as well as Kaplan-Meier estimates at three years. Technical success was defined as placement of both the main-body graft and successful stenting of target vessels in an “intent-to-treat” manner, further defined by the absence of an endoleak type I or III or graft obstruction, absence of the need to convert to open surgical repair, and survival > 24 hours.61 Outcomes regarding target vessels were reported as freedom from branch instability, which include occlusion, stenosis, and stent migration or fracture. All patients underwent computed tomography (CT) imaging with 1-mm sections. Aneurysms were classified as follows: infrarenal AAA, infrarenal AAA with conical neck, juxtarenal AAA (defined as an aneurysm extending up to but not involving the renal arteries), suprarenal AAA (defined as an aneurysm that extends up to the superior mesenteric artery, involving one or both renal arteries), or thoracoabdominal abdominal aortic aneurysm (TAAA), according to the Crawford-Safi classification.60 Data regarding baseline characteristics included age (years), gender, body mass index (BMI)(kg/m2) and aneurysm diameter (mm). Comorbidity data included preoperative existence of chronic obstructive pulmonary disease (COPD), diabetes mellitus, congestive heart failure and chronic renal insufficiency (creatinine > 150 µmol/l or renal replacement therapy). 21.

(22) For Paper IV, the FEVAR and BEVAR procedures were categorised by the number of target vessels intended for catheterisation, i.e. two, three or four, denoted as 2C, 3C, and 4C, respectively. There were ten procedural variables, which were categorised as radiologic-related, perioperative, or postoperative variables. Radiologic-related variables included: accumulated skin dose (Gy), contrast volume (ml), fluoroscopy duration (minutes), number of angiography series, and radiation/dose area product ((DAP) (Gy·cm2)). Perioperative variables included: blood loss (ml), anaesthesiology duration (minutes), and procedure duration (minutes). Postoperative variables included: post-operative intensive unit care (ICU) length-of-stay (days), and total hospital admission (days).. Paper V Patients and Procedures All patients admitted and surgically treated for a Stanford type A aortic dissection or intramural hematoma from 2004 – 2015 at Uppsala University hospital were included. Baseline demographics and comorbidities were collected, as well as the Debakey aortic dissection classification, the presence of connective tissue disease, bicuspid aortic valve, and bovine aortic arch variant. Finally, the living status of all patients as of April 01, 2019 was obtained. Data from the primary surgical procedures provided information regarding potential concomitant aortic valve replacement, coronary artery bypass grafting (CABG), in addition to whether or not a clamp for the distal aortic anastomosis was implemented. Any bypass to one of the supraaortic vessels or primary aortic arch repair was also noted. Imaging evaluation The most recent follow-up CT images for all patients were evaluated using the post-processing software 3mensio Vascular (3mensio Medical Imaging Bilthoven, The Netherlands). For patients who underwent any form of aortic arch repair since the primary procedure, the most recent image prior to the reoperation was used. Appropriate centre-, outer-, and inner-line measurements were obtained for the ascending aorta, defined as the distance from the Innominate artery (IA) to either the most proximal edge of prosthetic material or the most distal coronary artery, Figure 2. The maximum diameter was measured for the ascending and proximal descending aorta, as well as the aortic arch. Lengths and diameters were then obtained for the IA, the right common carotid (RCC), and Left Common Carotid (LCC) arteries. Dissection, as well as tortuosity and severe calcification, for these vessels were also noted. In cases of a bovine. 22.

(23) aortic arch, the length and diameters of this variant were measured, in addition to the angulation of RCC and LCC take-off vessels.. Figure 2. Demonstrated centre-line measurement from the level of the most distal coronary artery to the Innominate artery.. Endovascular Arch Inner-Branched Stentgraft Designed and manufactured by Cook Medical (Bloomington, IN, USA), the endovascular arch inner-branched stentgraft (AIBS) is available with either one or two proximal sealing stents, as well as internal branches for their respective supraaortic vessels (Figure 3).. 23.

(24) Figure 3. The arch inner-branched stentgraft viewed laterally (A), from within with internal branches (B), and from above with external markers beside the diamondshaped openings. Copyright permission from AME Publishing Company.. The graft requires a suitable and uniform ascending aorta with a diameter no greater than 38 mm, while the innominate artery (IA) and LCC artery should be no greater than 18 mm. Previous reports have indicated a minimal uniform sealing zone of 4.0 cm in the ascending aorta. Significant kinking of a previous implanted aortic graft, albeit subjective, is also a contraindication. When sealing lengths and angulation are concering, minimum outer and inner curve lengths of 45mm and 24mm, respectively, are mandated by the stentgraft manufacturers. Although previous reports have suggested allowances for the presence of a mechanical aortic valve, it is noted as a contraindication in this analysis. See Table 3 for the detailed criteria. Table 3. Technical criteria regarding eligibility for the arch inner-branched stent graft.* • • • • • •. Ascending aorta diameter ≤ 38 mm. Uniform ascending aorta with no significant angulation/kinking. Sealing zone in the ascending aorta with true, centre-line length ≥ 40 mm and/or outer-curve length ≥ 45 mm, inner-curve length ≥ 24mm. Suitable innominate and left common carotid artery landing zone with diameters ≤ 18 mm. Iliac artery access accommodating a minimum 22 French sheath. Native or biological aortic valve, i.e., mechanical aortic valve contraindicated.. *Written correspondence with Cook Medical (Bloomington, IN, USA) custom aortic stentgraft representatives.. 24.

(25) Outcomes The primary outcome was survival ≥ 1 year and fulfillment of the technical criteria for the currently available endovascular arch inner-branched stent graft given above. Ineligibility was noted for one or more of the following issues: ascending aorta diameter, ascending aorta length, prosthetic kinking, supraaortic landing zone suitability, presence of a mechanical valve, severe aortic valve insufficiency and root dilatation, or a combination thereof. It was also noted whether adjunct procedures would be necessary for eligibility, which were defined but not limited to either an interpositional graft placement in the LCC or RCC, or supraaortical deviation, i.e., carotid-subclavian bypass, supplemented with bridging stentgraft extension.. Statistics Data management and statistical analysis were performed using SPSS versions 23.0 and 24.0 (SPSS, Inc. Armonk, NY, USA: IBM Corp.), as well as Stata, version 14.2 (StatCorp. 2015. Stata Statistical Software: Release 14. College Station, TX, USA: StataCorp LP.) Continuous data for Papers I and II are presented with mean values and 95% confidence intervals (CIs), and compared with t-tests. Rates are presented as percentages with 95% CIs. Missing data were excluded. Comparison of rates between the two periods were tested with the chi-square test. Odds ratios were calculated for perioperative mortality using logistic regression, adjusting for the previously mentioned covariates. For Paper I, high and low volume centres were determined by ranking the participating centres based on the average number of procedures per year. Four quartiles were formed, while the top and bottom quartiles were considered high or low volume centres accordingly. For Paper II, volume per centre was determined by ranking the participating centres based on the average number of procedures for the four-year period. Five quintiles were formed (QI, highest-QV, lowest). Lastly, funnel plots were created using upper and lower CIs (95% and 99.8%) from the calculated mean perioperative mortalities, where values from each centre were displayed in the form of a scatter plot. For Papers III and IV, continuous data are presented as mean values ± standard deviation or median values with range or interquartile range (IQR). These were compared in Paper III using either t-tests or Mann-Whitney U tests. In Paper IV, ANOVA tests with post-hoc Bonferroni comparisons and Kruskal-Wallis tests were employed because of multiple groups. Categorical variables are reported as absolute numbers and compared using chi-square. 25.

(26) tests. Data on survival and target vessel instability were analyzed using Kaplan-Meier curves and compared using the log-rank test. To assess the relationships between the complex repairs and procedural variables in Paper IV, the one-tailed Pearson correlation analysis was used. Correlation was considered weak for a coefficient value of 0.1 ≤ r < 0.3, moderate for 0.3 ≤ r < 0.5, and strong for coefficient values r ≥ 0.5. The associated regression coefficient indicates the slope of the best fitting linear prediction and was used to calculate the predicted values for the linear increase in a variable with the increased number of catheterisations. The hypothesis for the statistical tests was that the slope was equal to zero. Before the analysis, Cook’s distance was calculated for all data points for each variable. Accordingly, outlying data points with a distance larger than 4/n (n = number of data points) were removed from that specific variable analysis and linear regression. Similar to Papers I and II, continuous data from Paper V are presented as mean values with 95% CIs. The categorical data are presented as absolute numbers (%) and compared using the Chi-squared test. Estimated survival was calculated using a Kaplan-Meier curves. Estimates of both death and aortic reoperation, as mutual competing risks, were obtained using a subdistribution hazard function on a cumulative incidence curve. The conventional p-value of 0.05 was chosen for statistical significance with the exception of Paper II, for which a significant p-value of 0.01 was chosen to correct for multiple comparisons.. 26.

(27) Ethical Considerations The data from Papers I and II were anonymized and collated at a single centre. The analyses were retrospective for both papers and approved by the individual participating national vascular surgery registries, and by the ethical review board, if needed, according to the ethical review process of each specific country. The analysis of the Swedish data was approved by the ethical review board of the Uppsala region and by the Swedvasc committee. The data from Papers III-V were also evaluated retrospectively. These retrospective cohort studies were approved by the local ethical review board.. 27.

(28) Results Paper I There were 34,375 patients from 2005-2009 and 48,878 patients from 20102013 included for the analysis. Over the two periods, EVAR increased from 44.3% to 60.6%, p<0.0001. These figures for each country are detailed in Figure 4. 100% 2005-2009. 2010-2013. 50% 68. 34. 28. nl. 29 32. 50 49 37. 61 44. 17. al. ar m. 44. k an d G er m a H ny un ga r Ic y el an d It al y N ew Ze al N and or w a Sw y ed en Sw itz er la nd U K. 15. 61. 57. 52. Fi. en D. us. tr. al. ia. 24. A. 54 49. 46 49. To t. 74 56. Figure 4. Percentage of EVAR for intact AAA repair used between two period overall, as well as for individual Vascunet participating countries.. Overall, perioperative mortality fell from 3.0% to 2.4%, p<0.0001. For EVAR, this rate fell from 1.5% to 1.1%, p<0.0001, while for open repair, this rate increased from 3.9% to 4.4%, p=.008, Figure 5. Moreover, the percentage of octogenarians, i.e., patients ≥ 80 years old, increased from 18.5% (95% CI, 18.1-18.9) to 23.1% (95% CI, 22.7-23.5), p<0.0001. Finally, the percentage of female patients decreased from 15.5% (95% CI, 15.0-15.9) to 14.1% (95% CI, 13.8-14.5), p<0.0001. Detailed perioperative mortalities for various subgroups for the latter period are given in Figure 6.. 28.

(29) Perioperative Mortality. 5.0%. 2005-2009. 2010-2013. 2.5% 4.4. 3.9 3.0 2.4 1.5. Overall. 1.1 EVAR. Open. Figure 5. Rates of perioperative mortality in the Vascunet registry between two time periods overall, as well as for EVAR and OAR.63. 29.

(30) EVAR. 10%. Open Surgery 9.7. 7.5% 6.0. 5%. 4.6. 4.0. 3.6. 0.7. ≥. 5. 5. 80. 80. en. 0.4. 1.1. 0.6. 1.3. ♂. A ge ≥. en M. nt A ll Pa tie. 1.1. 0.7. 3.4. 3.0. 2.5. < 5. 5 cm <8 0, A A A ♀ < < 5. 80 5 cm ,A A A ♂ < ≥ 5. 80 5 cm ,A A A ♀ < ≥ 5. 80 5 cm ,A A A < 5. 5 cm. 1.8. 0.9. A ge <. 1.0. s. 0%. 3.1. cm. 1.9. A A A. 2.5%. A A A. 4.4. W om. Perioperative Mortality. 9.5. Figure 6. Perioperative mortalities for various subgroups for both EVAR and open surgical repair for the period 2010-2013.63. The comparison of high and low volume centres for open repair and EVAR is detailed in Table 4. Table 4. Comparison of high- and low-volume centre perioperative mortalities for EVAR and OAR for 2005-2009 and 2010-2013.. EVAR High Volume Low Volume OAR High Volume Low Volume. 2005-2009. 2010-2013. p. 1.8% 1.3%. 1.1% 1.2%. 0.02 0.56. 3.4% 4.3%. 3.4% 5.4%. 0.86 0.03. EVAR: Endovascular aortic repair OAR: Open aortic repair. The logistic regression analysis, after correcting for age and patient sex, revealed a reduced perioperative mortality odds ratio when comparing the latter against the former period, OR=0.71 (95% CI, 0.64-0.78), p<0.0001. The OR for postoperative death after EVAR was significantly less, OR=0.59 (95% CI, 0.48-0.73), p<0.0001, while for open repair, it was significantly increased, OR=1.17 (95% CI, 1.03-1.30), p=0.01.. Paper II There were 9,273 patients included. EVAR was performed for 23.1% (95% CI, 22.3-24.0) of all RAAA procedures. The mean AAA size at the time of 30.

(31) operation was 7.6 cm (95% CI, 7.5-7.6). Of these aneurysms, 10.7% (95% CI, 10.0-11.4) were less than 5.5 cm. Table 5 includes a comparison of patient demographics and comorbidities between the EVAR and open repair patient cohorts. Table 5. Patient characteristics and comorbidities for EVAR and OAR.64 EVAR. OAR. p. Mean age, years (95% CI). 76.5 (76.1-76.9). 74.1 (73.9-74.4). <0.001. Women, % (95% CI). 18.4 (16.6-20.2). 16.9 (16.0-17.9). 0.15. 7.1 (7.0-7.2). 7.7 (7.7-7.8). <0.001. Ischemic Heart Disease, % (95% CI). 41.6 (39.5-43.8). 28.2 (27.2-29.3). <0.001. Pulmonary Disease, % (95% CI). 22.9 (20.7-25.0). 16.6 (15.6-17.5). <0.001. Diabetes Mellitus, % (95% CI). 11.9 (10.3-13.4). 11.0 (10.2-11.8). 0.30. 6.7 (5.2-8.2). 6.6 (5.9-7.3). 0.95. Mean AAA diameter, cm (95% CI). Cerebrovascular Disease, % (95% CI) AAA: Abdominal aortic aneurysm CI: Confidence interval EVAR: Endovascular aortic repair OAR: Open aortic repair. Overall perioperative mortality was 28.8% (95%CI, 27.9-29.8). For female patients, it was 32.3% (95% CI, 29.9-34.9), compared with male patients at 27.1% (95% CI, 26.1-28.2), p<0.0001. After adjusting for the above-mentioned comorbidities, the perioperative OR revealed no statistical difference between men and women, OR=1.03 (95% CI, 0.86-1.23), p=0.78. Perioperative mortality was 32.1% (95 CI%, 31.0-33.2) for OAR and 17.9% (95% CI 16.3-19.6) for EVAR. The adjusted OR for EVAR against OAR was 0.38 (95% CI, 0.31-0.47), p<0.0001. The data for each country and procedure are given in Table 6.. 31.

(32) Table 6. Overall and per procedure perioperative RAAA mortality rates for individual countries, 2010-2013.64 % EVAR. Overall mortality, %. EVAR mortality, %. OAR mortality, %. Total. 23.1. 28.8. 17.9. 32.1. Australia. 39.8. 18.4. 9.2. 24.5. Denmark. 5.1. 25.1. 10.5. 25.9. Finland. 9.9. 23.7. *. 26.3. Germany. 31.2. 35.5. 24.2. 40.6. Hungary. 7.5. 33.2. 7.1. 35.3. Iceland. 19.0. 19.1. *. 23.5. New Zealand. 10.9. 28.2. 16.7. 29.6. Norway. 11.7. 21.3. 15.4. 22.0. Sweden. 29.3. 29.6. 23.4. 32.2. Switzerland. 24.9. 22.8. 19.1. 24.0. United Kingdom. 18.0. 32.6. 20.3. 35.3. EVAR: Endovascular aortic repair OAR: Open aortic repair RAAA: Ruptured abdominal aortic aneurysm *There were no deaths in these cohorts.. The overall perioperative mortality was lower in centers performing EVAR for the majority of ruptures, i.e., primary EVAR (EVAR(p)) centres (23.0%, 95% CI, 20.6-25.4) compared to primary open aortic repair (OAR(p)) centres (29.7%, 95% CI, 28.6-30.8), p<0.001, Figure 7. The odds ratio for perioperative mortality was significantly lower for EVAR(p) centres, OR= 0.60 (95% CI, 0.46-0.78), p<0.001.. 32.

(33)

(34) Number of patients: 6816. 15.8% EVAR Open Repair. 84.2%. Perioperative Mortality Overall: 29.7% OAR: 32.1% EVAR: 16.9%.

(35) Number of Patients: 1217 Perioperative Mortality. 36.2%. EVAR. 63.8%. Open Repair. Overall: 23.0% OAR: 32.1% EVAR: 17.9%. Figure 7. Comparison of RAAA perioperative mortalities for primary OAR centres and primary EVAR centres, 2010-2013. The percentage of EVARs and open repairs performed are given within the figures.. Recognizing that 50% may be arbitrary as a demarcation for predominant strategy, a sensitivity analysis was carried out, reducing the percentage of RAAA patients treated with EVAR for EVAR(p) to ≥40%; the perioperative mortality was 20.2% (16.6-23.9) for EVAR(p) centres and 34.0% (31.3-36.7) for OAR(p) centres, p<0.001. An analysis of EVAR and OAR within centres of either EVAR or OAR as a predominant strategy yielded no differences of significance: the perioperative mortality for EVAR in EVAR(p) centres was 17.9% (95% CI, 15.2-20.6) and 16.9% (95% CI, 14.6-19.1) for OAR(p), p=0.56. For open aortic repair, the perioperative mortality was 32.1% (95% CI, 27.7-36.5) in EVAR(p) centres and 32.1% (95% CI, 30.9-33.3) in OAR(p) centres, p=0.99. 33.

(36) The relationship of EVAR as a primary strategy for RAAAs and intact AAAs gave the following: 95.4% (95% CI, 90.0-100.0) of the EVAR(p) centres used a primary strategy of EVAR for elective AAA repair, whereas only 58.2% (95% CI, 53.2-63.3) of the OAR(p) centres used EVAR as a primary strategy in treating their ruptures, p<0.0001. Alternatively, from the patients’ perspective, 82.0% (95% CI, 77.2-86.8) of the elective patients in EVAR(p) centres were treated with EVAR, while 50.7% (95% CI, 47.8-53.6) of the elective patients in OAR(p) centres were treated with EVAR, p<0.0001. There were 442 identified centres. None of the 13 highest volume centres (QI, >22 repairs per year) were EVAR(p). The greatest percentage of EVAR (30.6%) was performed in QII. The perioperative mortality was lowest in the highest volume centres, 23.3% (21.2-25.4), p<0.001. Perioperative mortality for OAR was significantly lower in the high volume centres, 25.3% (23.027.6), p<0.001. Despite the variation in perioperative mortality in EVAR between quintiles, there was no statistical differences between them, p=0.07. The relationships between centre volume and perioperative mortality for either EVAR or open repair are demonstrated in the form of funnel plots in Figures 8a and b.. Figure 8a. Funnel plot depicting number of open AAA repairs per centre and centre perioperative mortality for 2010-2013. Individual centres are represented by the black dots. The interrupted lines represent the 95% and 99.8% confidence intervals, while the target is the overall mean perioperative mortality.64. 34.

(37) Figure 8b. Funnel plot depicting number of EVARs per centre and centre perioperative mortality for 2010-2013. Individual centres are represented by the black dots.The interrupted lines represent the 95% and 99.8% confidence intervals, while the target is the overall mean perioperative mortality.64. Paper III There were 71 consecutive patients included for analysis. There were 40 juxta/pararenal AAAs (J/PAAA) and 31 patients with thoracoabdominal aortic aneurysms (TAAA), with the following distribution in Figure 9.. 35.

(38) Figure 9. Distribution of juxta/pararenal abdominal aortic and thoracoabdominal aortic aneurysms.. No patients were excluded for anatomical or technical reasons. There were 49 men and 22 women, and the median age was 71.9 years (range, 44-83). The median follow-up regarding patient living status was 25.0 (IQR, 14-39) months, while the median follow-up for target vessel imaging was 17.0 (IQR, 5-32) months. Two patients refused follow-up imaging. Of the 71 procedures, 47 were FEVARs (including 2 physician-modified fenestrated grafts) and 24 were BEVARs. There were no ruptures while awaiting treatment. The median operative time was 432 (IQR, 325-557) minutes for FEVAR procedures and 476.5 (IQR, 408.3-601.5) minutes for BEVAR procedures. Of the 71 patients, 21 (29.6%) had previously undergone aortic treatment. There were 4 cases of ruptured TAAAs. Mean aneurysm diameter at the time of treatment was 6.1±0.8 cm. Technical success was determined as follows: main-body stentgrafts were deployed successfully in 69 patients. Regarding target vessel technical success, there were 205/208 (98.6%) vessels successfully revascularized: 51/51 superior mesenteric arteries (SMA), 27/27 celiac arteries, and 127/130 renal arteries (left and right in one patient). Two failed renal vessel catheterizations were in one of the above-mentioned main-body failures. The other failed vessel catheterization was a severely stenosed left renal artery in a FEVAR procedure performed for a juxta/pararenal AAA; revascularization was later performed with open surgery. Combined with the above graft deployment failures, the technical success rate was 68/71 patients (95.7%).. 36.

(39) The 30-day mortality was 2 (2.8%). One patient died of multi-organ failure (juxta/pararenal AAA/FEVAR); the second patient (TAAA/BEVAR) died out-of-hospital, of sudden death and unknown causes. The 90-day mortality was 7 (9.9%). In addition to the two deaths within 30 days, the remaining were all procedural related, in-hospital, and a result of multi-organ failure, including permanent spinal ischemia (n=2), mesenteric ischemia and sepsis (n=1), and heparin-induced thrombocytopenia (n=1). Over the entire study period there were 15 deaths, comprised of 10 TAAAs (7 BEVARs and 3 FEVARs) and 5 juxta/pararenal AAAs (all FEVARs). Six of these deaths were a result of multi-organ failure; five were of unknown causes; one was due to bladder cancer; one patient died after a hemorrhagic stroke, and one died of worsening dementia and respiratory failure. Finally, one patient, whose aneurysm was unsuccessfully excluded during the primary operation, returned 25 months after the failed procedure with a rupture and died after refusal of treatment. There were otherwise no late aneurysm ruptures following successful treatment. The estimated combined survival at 3 years was 77.9% ± 5.6%. These estimates are stratified for juxta/pararenal AAAs, at 90.9% ± 5.2%, and for TAAAs at 60.7% ± 10.3%, log-rank P=0.01 in Figure 10.. Figure 10. Kaplan-Meier survival estimates for juxta/pararenal abdominal aortic aneurysm (AAA) and thoracoabdominal aortic aneurysm (TAAA) patients.65. 37.

(40) Branch instability, as defined above, was identified in nine cases during follow-up (5 in BEVAR procedures and 4 in FEVAR procedures) in 6 patients. Three of these were vessel occlusions (3 right renal arteries), all in patients who underwent BEVAR procedures (2 TAAAs, 1 juxta/pararenal AAA). Only one of these patients, who preoperatively had one functional kidney, underwent reintervention with thrombolysis and stent relining. The other two renal occlusions were observed at follow-up imaging at two months and two years, respectively, and appeared to have no clinical significance and were thus not treated. Of the nine unstable branches, one vessel (left renal artery) in a patient treated with BEVAR for a juxta/pararenal AAA, was identified with a stenosis at 30 days. At one year, there were a total of seven unstable branches (5 renal arteries and 2 SMAs (4 FEVARs and 3 BEVARs)). Two of these were the above-mentioned renal artery occlusions. There was one stent migration (FEVAR) at one year within the left renal artery that required supplementary selfexpanding stent placement. The remaining cases were stenoses and underwent percutaneous transluminal angioplasty (PTA) with supplementary stent placement. There were seven target vessels which underwent reintervention in 5/71 patients (7.0%). Five of the seven (71.4%) reinterventions were performed on renal arteries. The three-year Kaplan-Meier estimate for freedom from branch instability was 92.7% ± 2.5% overall. These rates are stratified according to either FEVAR or BEVAR in Figure 11, specifically, 88.6% ±6.4% for branched bridging stent grafts, and 94.6% ± 2.4% for fenestrated bridging stent grafts, log-rank P=0.28.. 38.

(41) Figure 11. Kaplan-Meier estimates of freedom from branch instability for branched and fenestrated bridging stent grafts among those surviving 90 days and available for follow-up imaging.65. Paper IV After exclusion, there were 63 EVAR, 40 FEVAR, and 22 BEVAR procedures available for analysis. Baseline patient characteristics and comorbidities were similar between the conventional EVAR and complex procedure groups. The complex procedures included 23 two-vessel, 20 three-vessel, and 19 four-vessel catheterizations. Results from the previously stated variables based on complexity are given in Table 7.. 39.

(42) Table 7. Radiologic, peri-, and post-operative variables for both standard EVAR and complex aneurysm repair, based on the number of vessels catheterized. The presented results are median values with the exception of the normally distributed anaesthesiology duration, which are given as mean values.66. EVAR Radiologic variables Accumu0.97 lated skin (0.03-7.19) dose, Gy (range). 2C. 3C. 4C. 2.79 (1.32-18.67). 4.62 (1.34-10.86). 7.63 (1.64-12.47). 29.5 (8-72). 35.0 (14-115). 49.0 (23-132). 77.5 (28-148). Fluoroscopy duration, min. (range). 24.8 (10.1-102.8). 71.7 (37.2-237.8). 89.0 (52.2-174.9). 125.7 (91.2-289.9). Number of angiography series (range). 10.0 (3-31). 14.5 (7-53). 25.5 (8-47). 31.5 (18-46). 152 (18-1320). 196 (96-793). 432 (113-893). 491 (204-939). 1 100 (150-5 500). 1 400 (100-5 500). 1 900 (650-4900). 181 (60-405). 340 (180-645). 458 (255-750). 628 (510-885). 126 (70-349). 359 (250-724). 433 (213-900). 555 (383-752). Contrast volume, g (range). DAP, Gy·cm2 (range). Perioperative variables Blood loss, 200 ml (0-1500) (range) Anesthesiology duration, min.a (range) Procedure duration, min. (range). 40.

(43) Postoperative variables Days in 0 post-opera(0-8) tive ICU (range) Total number of days at hospital (range). 4.0 (1-27). 0 (0-6). 0 (0-18). 2.5 (0-43). 5.0 (2-26). 7.0 (2-40). 13.0 (4-46). DAP: dose-area product EVAR: endovascular aortic repair Gy: Gray ICU: intensive care unit 2C: 2-vessel catheterisation 3C: 3-vessel catheterisation 4C: 4-vessel catheterisation a Normally distributed, mean values are presented.. The following strong linear relationships between the number of branch vessel catheterisations and variables were identified: accumulated skin dose (r=0.504), contrast volume (r=0.652), fluoroscopy duration (r=0.598), number of angiography series (r=0.650), anaesthesiology duration (r=0.742), procedure duration (r=0.554) and total length of stay (r=0.533). The correlations for blood loss (r=0.346) and DAP/radiation dose (r=.479) were moderate, while the correlation for days in the post-operative ICU (r=0.283) was weak.. Paper V There were 198 patients identified, 124 (62.6%) men and 74 (37.4%) women. The mean age at the time of operation was 61.4 years (95% CI, 59.8-62.9). There were six patients with Marfan syndrome. The 30-day and 1-year mortality were 16.2% (n=32) and 19.2% (n=38), respectively. The estimated 10year survival was 55.0% (95% CI, 45.6%-63.5%). There were 26 patients with no post-operative follow-up imaging, and four patients of the original cohort underwent aortic arch repair at the primary procedure. In addition to the 38 patients who died within one year, the remaining available cohort for analysis was 129 patients. The mean duration to the most recent CT imaging was 47.8 months (95% CI, 40.3-55.3). The primary outcome was met in 89 (69.0%) patients. The most common cause for ineligibility was the presence of a mechanical aortic valve (n=16, 40%), followed by inadequate landing-zone seal (n=12, 30.0%). No patients were ineligible due to supraaortic vessel pathology, although 18 (20.2%) of 41.

(44) the 89 patients would require an adjunct procedure, six of which were bilateral and 12 unilateral. Ineligible AIBS patients were younger and had shorter ascending aortic seal lengths. Distal clamping without circulatory arrest was also more often employed among these patients at the time of their primary operation, Table 8. Table 8. Patient and procedural data for patients found eligible or ineligible for an endovascular arch inner-branched stentgraft.. Eligible AIBS candidates (n=89). Ineligible AIBS candidates (n=40). p. 60.2 (58.0-62.4). 55.7 (52.0-59.3). .027*. 56 (62.9) /33 (37.1). 26 (65.0) /14 (35.0). .820. 68 (76.4%) 18 (20.2%) 3 (3.4%). 36 (90.0%) 3 (7.5%) 1 (2.5%). .071 .070 .792. Bovine Trunk (%). 7 (7.9%). 6 (15.0%). .213. CTD (%). 5 (5.6%). 3 (7.5%). .682. Ascending aortic seal length, mm (95% CI). 48.3 (46.4-50.1). 38.2 (34.0-42.3). <.001*. Ascending aortic diameter, mm (95% CI). 32.2 (31.5-32.9). 32.7 (31.6-33.8). .410 *. Circulatory Arrest, no distal clamp (%). 79 (88.8%). 29 (72.5%). .021. CABG or coronary reimplantation (%). 2 (2.2%). 13 (32.5%). <.001. Mechanical valve (%). 0. 20 (50.0%). <.001. Biological valve (%). 8 (9.0%). 3 (7.5%). .779. Mean Age (95% CI) Patient Sex Male (%) /Female (%) Pathology Type I AAD (%) Type II AAD (%) IMH (%). Reason for ineligibility (%). 42. 16 (40.0%).

(45) Mechanical Valve only Inadequate seal only AI/Root dilatation Diameter Kink Combination. 12 (30.0%) 2 (5.0%) 2 (5.0%) 1 (2.5%) 7 (17.5%). AAD: Stanford type A aortic dissection AI: Aortic insufficiency AIBS: Arch Inner-Branched Stentgraft CABG: Coronary artery bypass graft CTD: Connective tissue disease IMH: Intramural hematoma *: Compared using t-tests; all other tests were performed using Chi-squared tests.. Of the 129 followed patients, 19 (14.7%) developed aortic arch/proximal descending aortic dilatation ≥ 55m, while 14 (73.7%) would be considered AIBS eligible. There were 24 patients (18.6%) who underwent aortic reoperation at a mean follow-up of 7.1 years (95% CI, 6.6-7.7). The subdistribution hazards model in Figure 12, with death as the competing risk, yields an aortic reoperation risk of 14.3% (95.0% CI, 9.1-20.5). There were 13 arch repairs (68.4%) performed, of which two were treated with the endovascular AIBS (in the more latter part of the study period). The remaining 11 patients underwent open aortic repair for the following indications: aortic insufficiency (5), endocarditis (4), pseudoaneurysm (2).. 43.

(46) Figure 12. The cumulative incidence for both reoperation and survival, with death and reoperation as mutual competing risks.. 44.

(47) General Discussion Clearly, EVAR is here to stay. Its increasing use in aortic repair has been matched by small, yet impressive improvements. The reduction from an already low perioperative mortality of 1.5% to 1.1% in intact AAA repair, for example, is a remarkable finding. There is a back side, however, as success can readily become its own worst enemy, often the case for technological advancements. The indications for treatment can blur: older and more frail patients are now offered treatment, as are patients with dubious anatomical allowances. Although studies have shown, particularly in the form of the randomized controlled EVAR-2 trial, that morbid patients do not benefit from EVAR or, in the form of the small aneurysm trials, that patients do not benefit from earlier EVAR intervention, the inclination to challenge aneurysmal disease with EVAR appears unabated.23,67,68 Does real world practice ignore the evidence, or are the purported successes too tempting to overlook? A straightforward answer to this question may actually be less important than what a deeper evaluation of the evidence and trends reveals. Randomized controlled trials have been instrumental in establishing guidelines of care, while findings from real-world registries have supplemented them with important revelations as to what works and what flounders. It is critical that the various studies are balanced against each other and interpreted carefully, not necessarily an easy task in the context of rapidly evolving technology. It has been the goal of this thesis to capture the contemporary status of aortic treatment, in light of previous research, and to provide a provocative framework for the challenges that modern treatment has already embarked upon. The uptake in EVAR comes as no surprise, yet the increase, as demonstrated in Figure 4 is remarkable on several fronts. Rates have increased universally, with the overall rate surpassing 50% over the 9-year period. Finland, notably, has more than tripled its use. This can somewhat be explained by the timing of the above-mentioned randomized controlled trials, demonstrating the merits of EVAR in intact AAAs, but this only partially answers the question as to why. Furthermore, one could ask whether there is a point of diminishing return, or should perhaps all patients be treated with EVAR? Beck et al have also recently reported on the international variations in the treatment of aneurysms.69 They found a significantly greater use of EVAR in countries for which reimbursement was provided for services in countries like USA or Australia, as opposed to countries where no financial incentives are awarded, such as the northern European countries. Moreover, the use EVAR correlated well with the percentage of small aneurysms and the percentage of octogenarians being treated. More than a quarter of the intact AAAs treated in Paper I were less than 5.5.cm, and the variation between countries was also marked, e.g., 9.2% in the UK and 46.1% in Germany. 45.

(48) Thus, a mollifying of the AAA-diameter thresholds, for whatever reasons, will increase the number of patients offered treatment. This trend should be appreciated with caution. Although, as shown in Figure 6, the perioperative mortality for small AAAs is slightly lower than it is for AAAs > 5.5cm, the risk of perioperative death was >3% after open repair, and 5-10% among female octogenarians undergoing open surgery for a small aneurysm. It should be recalled that the risk of rupture of aneurysms <5.5 cm was less than 1% in the small aneurysm randomized trials referred to above. Coupled to this discussion is the growing awareness that perioperative mortality following open repair is worsening. The above-reported change from 3.9% to 4.4% over time may be a result of several factors. First, the bias of selecting only the technically demanding cases for open surgery may play a role. Landry et al have reported on the increasing complexity of open repair in the EVAR era, with more frequent suprarenal aortic cross-clamping and longer operative times (see below, regarding complex aortic repair).70 It is also recognized that hostile infrarenal AAA necks have worse outcomes.71,72 The results following open repair and particular patient cohorts are placed in context in Table 9. Table 9. Perioperative intact AAA mortality following open surgery for various patient cohorts.. Schermerhorn et al73 Raval et al74 Patel et al75 Deery et al76. Year of Publication. Patient Cohort. 2015 2012 2011 2017. Medicare Octogenarians Complex AAAs Women. Open Aortic Repair Perioperative Mortality 5.2% 6.1% 5.7% 8.0%. Second, surgical outcomes are dependent on competency and routine, a well-established relationship of volume and outcomes, and the new generation of vascular surgeons may be bearing this out.77 The results regarding volume in Table 4 are compelling, suggesting that centres of high volume can maintain good results, and centralization may be warranted, particularly for AAA open repair. The topics raised above in the treatment of intact AAAs are not incongruent with those that merit attention in the treatment of RAAAs in Paper II. First, the overall mortality of 28.8% is in accordance with results reported from other recent studies, Table 10. Second and, again similar to other registry analyses, the mortality rates are lower for EVAR than they are for open repair. Of course, retrospective analyses of RAAAs have an unappealing tendency towards biases of patient selection, as EVAR often requires clinical and anatomic allowances for proper preoperative imaging and adequate aneurysm sealing. Vigilant of these biases, there are nonetheless several important attributes of the present analysis. 46.

(49) First, its international scope offers a glimpse of worldwide trends. There is an increasing capacity, or tendency, to treat ruptured patients with EVAR, which should not be understated. That is, this uptake requires complex restructuring of vascular surgery services, including referrals and centralizations (more below). The variation in the rates of EVAR for RAAAs in Table 6, e.g. 5.1% in Danmark and 29.3% in its neighboring country, Sweden, are noteworthy in regards to how different countries adapt to this. Table 10. Perioperative mortalities for either EVAR or OAR for ruptured abdominal aortic aneurysms from recent retrospective analyses. Authors. Year of Publication. Setting. Mani et al56 Chen et al78 Mohan et al79 Karthikesalingam et al80. 2011 2013 2013 2015. Taylor et al81. 2016. International Taiwan USA England Sweden New Zealand. Perioperative Mortality EVAR 20% 44% 26% 28% 21% 18%. OAR 33% 38% 39% 41% 31% 36%. EVAR: Endovascular aortic repair OAR: Open aortic repair. When comparing EVAR and open repair, it is interesting to note that EVAR patients were older and more heavily burdened with cardiac and pulmonary comorbidities. Incorporation of these and other covariates in a logistic regression revealed a perioperative odds ratio of 0.38 in favor of EVAR. This may address some, but certainly not all, of the concerns of selecting certain, potentially “better”, patients for EVAR. The analysis of outcomes based on the predominant strategy offers further insight. By deliberately dividing the cohort into their predominant treatment, one gains a sense of strategy, i.e., a sense that particular centres opt more often for one type of procedure, instead of choosing a procedure based on the stability or clinical fitness of the patient. Figure 7 depicts this, revealing that centres predominantly employing EVAR, and centres predominantly employing OAR, are equally competent in both EVAR and open repair. The superior results of EVAR(p) centres is a mere reflection of the greater proportion of EVAR procedures they performed, 64% versus 16%. Pursuant to the above-mentioned undertaking of adapting to endovascular strategies, there also appears to be a strong relationship between how patients are treated electively and acutely. Almost 100% of the EVAR(p) centres treated the majority of their elective patients with EVAR, whereas just under 60% of the OAR(p) centres treated the majority of their elective patients with EVAR. This can be interpreted in several ways. It may be ipso facto that acute provision of EVAR requires, or is explained by, routine utilization of EVAR in the elective setting, but many OAR(p) centres treat their elective patients 47.

(50) with EVAR, yet treat their acute patients with open repair. This may be an issue of institutional policy or an issue of individual surgeon preference. The question is: how long is this sustainable? This is an important topic for vascular centres to address in future reorganizations. Related to reorganization is the issue of centralization and patient-volume. The results show better outcomes, at least for open surgery, in centres of high volume, and this has been shown both in Paper I as well as in other large analyses looking at RAAA outcomes.82,83 The same cannot be said for EVAR, although a type II error may be at play. The funnel plots in Figure 8 also demonstrate this. Perioperative mortalities for open repair are more likely to lie closer or beneath the mean with increasing volume, whereas the opposite appears true for EVAR. This may reflect a referral pattern of more challenging EVAR patients to centres of expertise with higher volumes, but this cannot be ascertained from the available data. Underlying much of the analysis is, of course, the absence of data regarding turn-down rates. Exclusion of these patients can skew outcomes. Centres of higher volume may turn down fewer patients and, thus, potentially experience an increased rate of mortality for their efforts. The discussion of centralization and the establishment of centres of expertise is also pertinent in the treatment of anatomically complex aortic aneurysms. The embracement of a total endovascular approach for complex aneurysms, and the ability to document this undertaking, was the topic of Paper III. For open repair of thoracoabdominal aortic aneurysms, the commendable results from a high-volume centre could not be achieved in a wider setting.84 Juxtaposed with this is the rapid increase in complex endovascular repairs with good results, Table 11.85–87 Consideration of a total endovascular requires multiple metrics, and an analysis of outcomes, as given above, goes far to meet this need. The patients included were consecutively enrolled, included four ruptures, and none were excluded for anatomical or technical reasons. The technical success rate of 95.7% and 30-day mortality of 2.8% are favorable, as are the mid-term estimates from the Kaplan-Meir curves. Table 11. Technical success and mortality from recent studies regarding endovascular repair of complex aneurysms.. Verhoeven et al87 Martin-Gonzalez et al88 Eagleton et al89 Kristmundsson et al90. Year of Aneurysm Type Publication 2010 J/PAAA 2015 J/PAAA TAAA 2016 TAAA 2014 J/PAAA. J/AAA: Juxta/Pararenal abdominal aortic aneurysm TAAA: Thoracoabdominal aortic aneurysm. 48. Technical Success 87% 95% 99% 94% 91%. Mortality 5 year-93% 2 year-86% 2 year-95% 3 year-57% 5 year-60%.

(51) Technical success and mid-term results must, however, be appreciated along with other factors. These procedures are demanding, as testified by many of the strong correlations shown in Paper IV. Katsargyris and colleagues have argued that the increasing complexity has no technical nor clinical consequence, going further to suggest that the more complex procedure should be favored, given the presumption that a more comprehensive treatment would be more durable.91 This has been counterbalanced by Makaloski et al, who demonstrated an increased early all-cause mortality among octogenarians undergoing complex aneurysm repair.92 Anticipating an increase in elderly patients, they suggest that less complex procedures may be warranted, particularly when durability is less of a concern in older patients. It is important to point out, however that, just as age and patient comorbidities may evolve, the technical capacity to accommodate and assuage complexity will also improve. Procedures using fusion imaging were excluded from Paper IV, and these types of innovations have already made an impact.93 It is critical to continue to follow and document these changes in relationships of complexity and outcomes. The issue of treating patients with complex aneurysms has two other important and insidious elements. First, a recent study from the U.S. included over one thousand patients who underwent complex aneurysm repair, and they found an overall five- and seven-year survival of 46% and 30% respectively.94 Very few of the deaths were aneurysm or procedure-related, a stark gauge that these patients have severe underlying comorbidities. Second is the question of cost. A recent cost-effectiveness analysis by Michel et al, using data from the WINDOWS registry, found that costs were substantially higher for endovascular treatment than for open surgical repair, despite no differences in mortality.95,96 These two studies underscore aspects beyond feasibility and will need to be addressed moving forward. As stated previously, no randomized controlled trials exist for patients with complex aneurysms, and one hearkens to the comment by Greenberg et al, in their comparison of open and endovascular treatment, that it may be more appropriate to randomize these patients to intervention or no intervention.49 Nonetheless, the aspiration for healing the aorta with new technology has progressed, encroaching now on the ascending aorta, the realm of which Spear and his colleagues characterize as, “the zenith” of technical complexity.97 Initially introduced as case reports, stentgraft treatment of the ascending aorta was used in highly selected cases.53,98 There are several obstacles and concerns, not least of which include proximity to coronary vessels and valves, as well as deployment precision. Open surgery is the gold standard for much of ascending aortic pathology, particular dissections and aneurysms.99,100 Provided there is an adequate landing zone, it is felt that newer and specifically designed stentgrafts can be deployed with the added benefit of minimal inva-. 49.

(52) siveness. Moreover, if the patient had previously undergone an open procedure with prosthetic graft placement, this improves the suitability of stentgraft sealing while, at the same time, reduces the morbidity of a redo sternotomy. Early reports of this treatment have been limited but promising.51 In the largest and most recent analysis of 70 post-dissection patients treated with AIBS, Verscheure and colleagues report a technical success of 94.3% and a low combined mortality and stroke rate of 4.3%.52 There are three salient findings in Paper V, two of which confirm the already established characteristics of post-dissection patients. That is, many survive and many require a reoperation. The third important finding is that the majority of these patients are eligible for an endovascular repair, should the indication arise. If mechanical aortic valves and shorter landing zones can be accommodated, the eligibility would approach 100%. This is a remarkable attainment, given the infancy of this technology. Durability remains to be evaluated, and patient selection needs to be refined. This remark brings the issue almost full circle to some of the original questions of whether endovascular treatment was feasible, whether it was comparable, and ultimately whether it was superior to conventional treatment. Again, the success of EVAR has partially been its own flaw. Surgical triumphs tend to spawn new challenges, provide motivation for more complex procedures and dubious patient selection. The devices have changed and presumably improved over time, a persistent disparagement against older randomized trials, but newer devices do not necessarily confer better results. This is relevant in light of a recent analysis of the novel endovascular aneurysm sealing (EVAS) device, a stent graft designed to reduce the occurrence of endoleaks.101 The recent analysis from a single-centre found a dismal four-year freedom from failure rate of 42%, with eight secondary aortic ruptures.102 This is important for several reasons. First, newer devices are not per se an improvement because of their novelty. Second, appropriate use of registry data was eminently effective in uncovering these consequential findings . Last, and in light of the above-raised issue of newer technology and costs, new treatments, such as EVAR or F/BEVAR, are mandated with more than mere submissions of feasibility. Durability and cost-effectiveness are also needed in order for health care services to continue offering them. These concerns appear to have been weighed in the recent draft of the National Institute for Health and Care Excellence in the United Kingdom regarding AAA treatment.103 In it, they point out that, even if long-term benefits of EVAR could be achieved in place of open surgery, they could not counterbalance the costs that this would incur. This places a problematic, although not insurmountable, burden on the future of endovascular therapy in aneurysm treatment. Its successes must be used more judiciously and with greater acuity, in conjunction with both randomized and retrospective analyses, in order to improve patient and treatment selection.. 50.

No results found